This Title All WIREs
How to cite this WIREs title:
WIREs Clim Change
Impact Factor: 3.462

Stratospheric temperature trends: our evolving understanding

Full article on Wiley Online Library:   HTML PDF

Can't access this content? Tell your librarian.

We review the scientific literature since the 1960s to examine the evolution of modeling tools and observations that have advanced understanding of global stratospheric temperature changes. Observations show overall cooling of the stratosphere during the period for which they are available (since the late 1950s and late 1970s from radiosondes and satellites, respectively), interrupted by episodes of warming associated with volcanic eruptions, and superimposed on variations associated with the solar cycle. There has been little global mean temperature change since about 1995. The temporal and vertical structure of these variations are reasonably well explained by models that include changes in greenhouse gases, ozone, volcanic aerosols, and solar output, although there are significant uncertainties in the temperature observations and regarding the nature and influence of past changes in stratospheric water vapor. As a companion to a recent WIREs review of tropospheric temperature trends, this article identifies areas of commonality and contrast between the tropospheric and stratospheric trend literature. For example, the increased attention over time to radiosonde and satellite data quality has contributed to better characterization of uncertainty in observed trends both in the troposphere and in the lower stratosphere, and has highlighted the relative deficiency of attention to observations in the middle and upper stratosphere. In contrast to the relatively unchanging expectations of surface and tropospheric warming primarily induced by greenhouse gas increases, stratospheric temperature change expectations have arisen from experiments with a wider variety of model types, showing more complex trend patterns associated with a greater diversity of forcing agents. WIREs Clim Change 2011 2 592–616 DOI: 10.1002/wcc.125

Figure 1.

Percentage of expected radiosonde temperature reports received during October 2010 by the European Centre for Medium‐Range Weather Forecasts (ECMWF) for the (bottom to top) 50‐, 10‐, and 5‐hPa levels. A comparable map showing 700‐hPa reporting performance is shown by Thorne et al.1 [Figure courtesy of Antonio Garcia‐Mendez (ECMWF)]

[ Normal View | Magnified View ]
Figure 2.

Vertical sampling of satellite and radiosonde observations of stratospheric temperature. Left: vertical weighting functions for satellite Microwave Sounding Unit (MSU) and Stratospheric Sounding Unit (SSU) stratospheric temperature observations as a function of pressure (left axis) and height (right axis). The dashed line at about 27 km (30 hPa) indicates the typical maximum height of historical global radiosondes data coverage (Figure 1). Right: schematic of atmospheric vertical structure and its latitudinal variation. (Modified from Climate Change Science Program Synthesis and Assessment Product 1.14)

[ Normal View | Magnified View ]
Figure 3.

Smoothed global mean lower stratospheric temperature anomalies for 1958–2010 based on five radiosonde [Hadley Centre Atmospheric Temperatures (HadAT),10 Radiosonde Atmospheric Temperature Products for Assessing Climate (RATPAC),11 Iterative Universal Kriging (IUK),12 RAdiosonde OBservation COrrection using REanalyses (RAOBCORE),13 and Radiosonde Innovation Composite Homogenization (RICH)14] and three MSU [Remote Sensing Systems (RSS),15 University of Alabama in Huntsville (UAH),16 and NOAA Center for Satellite Applications and Research (STAR)17,18] datasets. Radiosonde data at different pressure levels have been averaged to correspond with the Microwave Sounding Unit (MSU) weighting function (Figure 2). The bottom trace is the mean of four of the five radiosonde datasets (excluding IUK, which does not extend beyond 2005). Anomalies are differences from 1979 to 1998 monthly mean values. Symbols indicate major volcanic eruptions with aerosols penetrating into the stratosphere in 1963 (Agung), 1982 (El Chichón), and 1991 (Pinatubo). Differences between individual datasets and the radiosonde mean are shown separately for the MSU (top) and radiosonde (middle) datasets. (Updated and modified from State of the Climate in 200819 and courtesy of Carl Mears, Remote Sensing Systems, and Katharine Willett, UK Met Office Hadley Centre)

[ Normal View | Magnified View ]
Figure 4.

Global temperature anomaly time series from Stratospheric Sounding Unit (SSU) data from the three SSU channels, as analyzed by Randel et al.21 and, for channels 25 and 26, as analyzed by Liu and Weng,29 and differences between them for channels 25 and 26. Symbols indicate major volcanic eruptions in 1982 (El Chichón) and 1991 (Pinatubo).

[ Normal View | Magnified View ]
Figure 5.

Modeled and observed global stratospheric temperature trend profiles. Left: simulated global mean temperature trends for approximately 1980–2000, based on an average of several models, due to changes in long‐lived greenhouse gases (LLGHGs) (including CO2, CH4, and N2O) and various ozone‐depleting substances (ODSs), changes in stratospheric ozone, and two different estimates (from balloon and satellite observations) of changes in stratospheric water vapor (Modified from Shine et al.85). Middle: simulated and observed 1979–2005 global mean temperature trends. Simulations are multimodel means, from seven chemistry‐climate models (CCMs), of trends due to changes in LLGHGs (including CO2, CH4, N2O, and associated changes in ozone and water vapor), ODSs [including chlorofluorocarbons (CFCs), hydrochlorofluorocarbons (HCFCs), and Halons and associated simulated changes in ozone, etc.], and natural forcings (volcanic aerosols and solar changes). Observations are from Stratospheric Sounding Unit (SSU) and Advanced Microwave Sounding Unit (AMSU)/Microwave Sounding Unit (MSU). Symbols are plotted at representative pressures for each satellite channel, and model trends, vertically weighted to correspond with these channels, are plotted at the same pressures (Modified from Gillett et al.86). Right: simulated and observed 1980–1999 near‐global (70°N–70°S) temperature trends. Simulations are by 18 CCMs; individual and multimodel mean results are shown. Observations are from MSU and SSU. Error bars on observed trends are 95% confidence intervals. (Modified from CCMVal Report69 and Forster et al.87)

[ Normal View | Magnified View ]
Figure 6.

Evolution of estimates of observed cooling trends in global mean lower stratospheric (LS) temperatures during the satellite era (since 1979), based on satellite Microwave Sounding Unit (MSU) (blue) and radiosonde (red) observations. Symbols show trends for 1979 to the year plotted, as reported in the literature, except for 1979–2009 trends which were calculated for this study. Radiosonde data are vertically weighted to correspond with the MSU LS layer (Figure 2). Blue line shows trends from the current (September 2009) version of University of Alabama in Huntsville (UAH) MSU data for each year. Differences between this line and the UAH published estimates (blue circles) illustrate the degree of change in the different versions of this dataset. See Figure 3 legend for dataset names and associated references.

[ Normal View | Magnified View ]
Figure 7.

Global mean temperature anomalies for 1979–2005 in the four vertical layers sampled by Microwave Sounding Unit lower stratosphere (MSU LS),15 Stratospheric Sounding Unit (SSU) 25, SSU 26, and SSU 27.21 The bottom trace is a combination (the average of channels 25 and 27 minus channel 26) of time series for the SSU layers from the models and as observed. See Figure 2 for the vertical layers. Observations are shown in black and simulations from eight models participating in Chemistry‐Climate Model Validation Activity (CCMVal)69 are shown in red.86 All anomalies are relative to the full 1979–2005 period.

[ Normal View | Magnified View ]
Figure 8.

Global map (left) and latitudinal variation (right) of Microwave Sounding Unit lower stratospheric (MSU LS) temperature trends during 1979–2010, based on Remote Sensing Systems (RSS) version 3.3 data. Error bars on zonal mean trends are 2‐σ estimates of internal data uncertainties.176 (Figure courtesy of Carl Mears, RSS)

[ Normal View | Magnified View ]
Figure 9.

Lower stratospheric (LS) temperature trends during 1979–2009 as a function of latitude and month from Microwave Sounding Unit (MSU) LS observations (top) and radiosonde 50‐hPa observations (bottom). Each panel shows a composite of trend estimates from different datasets, including the Remote Sensing Systems (RSS), University of Alabama in Huntsville (UAH), and NOAA Center for Satellite Applications and Research (STAR) Microwave Sounding Unit (MSU) datasets and the Hadley Centre Atmospheric Temperatures (HadAT), RAdiosonde OBservation COrrection using REanalyses (RAOBCORE), Radiosonde Atmospheric Temperature Products for Assessing Climate (RATPAC), and Radiosonde Innovation Composite Homogenization (RICH) radiosonde datasets. Areas without stippling have trends that are statistically significant at the 95% confidence level in all datasets in the composite and all are of the same sign. (Updated and modified from Fu et al.,177 Forster et al.178 and Free.179 Figure courtesy of Melissa Free, NOAA Air Resources Lab)

[ Normal View | Magnified View ]

Related Articles

Tropospheric temperature trends: history of an ongoing controversy

Browse by Topic

Paleoclimates and Current Trends > Modern Climate Change
blog comments powered by Disqus

Access to this WIREs title is by subscription only.

Recommend to Your
Librarian Now!

The latest WIREs articles in your inbox

Sign Up for Article Alerts

Twitter: earth_wise Follow us on Twitter

    Are you attending ICID 2014? Take a look back at 2013 with the Special Issue from Irrigation and Drainage @2014ICID http://t.co/LnzK6pZHr7
    Just over a month until the #GSA2014. We're excited and we hope you are too. Come to the Wiley stand for many surprises and freebies!